Remove flat: scripts/embed_modal.py

scripts/embed_modal.py

@@ -1,167 +0,0 @@
-"""Modal GPU job for CLIP embedding — ~10 min on A10G, no local storage.
-
-Mounts IMPACT from HF, runs CLIP ViT-L/14 on GPU, pushes parquet to HF Hub.
-
-Setup (one time):
-    pip install modal
-    modal setup          # authenticate
-    modal secret create hf-secret HF_TOKEN=hf_...
-
-Run:
-    modal run scripts/cloud/embed_modal.py
-
-Output: hf://datasets/midah/patent-wireframes/embeddings_2022_vitl14.parquet
-"""
-
-import io
-import os
-
-import modal
-
-# ── Modal image with all dependencies ────────────────────────────────────────
-
-image = (
-    modal.Image.debian_slim(python_version="3.11")
-    .pip_install(
-        "open_clip_torch",
-        "huggingface_hub",
-        "datasets",
-        "Pillow",
-        "pandas",
-        "numpy",
-        "tqdm",
-    )
-)
-
-app = modal.App("patent-clip-embed", image=image)
-
-# Secret holds HF_TOKEN
-hf_secret = modal.Secret.from_name("hf-secret")
-
-
-@app.function(
-    gpu="A10G",
-    timeout=3600,
-    secrets=[hf_secret],
-    memory=16384,
-)
-def embed_year(year: str = "2022", model_name: str = "ViT-L-14", pretrained: str = "openai"):
-    import ast
-    import csv
-    import zipfile
-
-    import numpy as np
-    import open_clip
-    import pandas as pd
-    import torch
-    from huggingface_hub import HfApi, hf_hub_download
-    from PIL import Image
-    from tqdm import tqdm
-
-    token = os.environ["HF_TOKEN"]
-    device = "cuda"
-
-    # Load CLIP model
-    print(f"Loading {model_name} ({pretrained}) on {device}...")
-    model, _, preprocess = open_clip.create_model_and_transforms(
-        model_name, pretrained=pretrained
-    )
-    model = model.to(device).eval()
-
-    # Download IMPACT CSV (small — just metadata)
-    csv_path = hf_hub_download(
-        repo_id="AI4Patents/IMPACT",
-        filename=f"{year}.csv",
-        repo_type="dataset",
-        token=token,
-    )
-
-    # Build list of (patent_id, figure_num, image_filename)
-    figures = []
-    with open(csv_path) as f:
-        for row in csv.DictReader(f):
-            try:
-                fnames = ast.literal_eval(row["file_names"])
-                pid = row["id"]
-                for i, fn in enumerate(fnames):
-                    figures.append({"patent_id": pid, "figure_num": i, "filename": fn})
-            except Exception:
-                pass
-
-    print(f"Total figures: {len(figures):,}")
-
-    # Download the zip (4.4GB — into Modal's /tmp, ephemeral)
-    print("Downloading IMPACT zip...")
-    zip_path = hf_hub_download(
-        repo_id="AI4Patents/IMPACT",
-        filename=f"{year}.zip",
-        repo_type="dataset",
-        token=token,
-        local_dir="/tmp/impact",
-    )
-
-    # Embed in batches
-    BATCH = 64
-    all_ids, all_vecs = [], []
-
-    def load_image(fn: str) -> Image.Image | None:
-        parts = fn.split("-D0")
-        if len(parts) < 2:
-            return None
-        inner_path = f"{year}/{parts[0]}/{fn}"
-        try:
-            with zipfile.ZipFile(zip_path) as z:
-                with z.open(inner_path) as f:
-                    return Image.open(io.BytesIO(f.read())).convert("RGB")
-        except Exception:
-            return None
-
-    batch_imgs, batch_ids = [], []
-
-    def flush_batch():
-        if not batch_imgs:
-            return
-        tensors = torch.stack([preprocess(im) for im in batch_imgs]).to(device)
-        with torch.no_grad():
-            feats = model.encode_image(tensors)
-            feats = feats / feats.norm(dim=-1, keepdim=True)
-        all_vecs.append(feats.cpu().numpy())
-        all_ids.extend(batch_ids)
-        batch_imgs.clear()
-        batch_ids.clear()
-
-    for fig in tqdm(figures, desc="Embedding"):
-        img = load_image(fig["filename"])
-        if img is None:
-            continue
-        batch_imgs.append(img)
-        pid = fig["patent_id"].lstrip("D").zfill(7)
-        batch_ids.append(f"D{pid}_{fig['figure_num']}")
-        if len(batch_imgs) >= BATCH:
-            flush_batch()
-    flush_batch()
-
-    vecs = np.vstack(all_vecs).astype(np.float32)
-    print(f"Embedded {len(all_ids):,} figures, shape {vecs.shape}")
-
-    # Save parquet and push to Hub
-    df = pd.DataFrame({"figure_id": all_ids, "embedding": list(vecs)})
-    out = f"/tmp/embeddings_{year}_{model_name.lower().replace('-','')}.parquet"
-    df.to_parquet(out, index=False)
-
-    out_file = f"embeddings_{year}_vitl14.parquet"
-    api = HfApi(token=token)
-    api.upload_file(
-        path_or_fileobj=out,
-        path_in_repo=out_file,
-        repo_id="midah/patent-wireframes",
-        repo_type="dataset",
-    )
-    print(f"Pushed → hf://datasets/midah/patent-wireframes/{out_file}")
-    return {"n_embedded": len(all_ids), "shape": list(vecs.shape)}
-
-
-@app.local_entrypoint()
-def main():
-    result = embed_year.remote("2022")
-    print("Done:", result)